Вы находитесь на странице: 1из 34

IJB Volume

5 Number
4 December
2001,
403 436
Accessing
bilingual
codeswitching
competence

403

Accessing bilingual code-switching


competence*
Almeida Jacqueline Toribio
The Pennsylvania State University

Acknowledgment*
Versions of this paper were presented at the 17th National Conference on Spanish in the United States at
Florida International University in March, 1999 and at the City University of New York Graduate Center in
October, 1999. A briefer version is to appear in a volume of selected papers from the former conference
(Toribio 2000e). I am grateful to the audiences of these gatherings for helpful comments. I would also like
to recognize the contributions of two anonymous readers for The International Journal of Bilingualism,
whose constructive criticisms have led to considerable improvements in this final version. All errors and
omissions are, of course, the sole responsibility of the author. This research was supported in part by a
fellowship from the National Endowment for the Humanities (FA- 34114- 96).

Abstract

Key words

The present work is inspired by an interest in the syntactic regularities that underlie
language alternations in Spanish-English bilingual speech, and the methodologies
that may prove most reliable and informative in this exploration. Accordingly,
it attends to the conceptual and methodological issues that must be addressed
and surmounted, and taking account of these concerns, presents new methodologies for gathering codeswitching data. The robust findings attest to the
validity of the methodologies and of the elicited data, affording a valuable
source of facts for further research.

codeswitching
competence
syntax

1 Introduction
The present work is inspired by an interest in the syntactic regularities that underlie language
alternations in Spanish-English bilingual speech, and the methodologies that may prove most
reliable and informative in this exploration. To that end, it critically surveys the research
literature, attending to the conceptual and methodological issues that must be addressed and
surmounted in the study of codeswitching competence. Taking account of these concerns,
the work presents new methodologies for gathering codeswitching data, examining permissible and unacceptable language alternations in the reading, recounting, and writing of
codeswitched narratives by 10 Spanish-English bilinguals, and the potential differential
status of codeswitching across these controlled and naturalistic tasks. The findings attest to
the validity of the methodologies and of the elicited data, which converge in revealing bilinguals strong sensitivity to syntactic well-formedness in all three conditions. The study thus
affords a valuable source of facts for further research into codeswitching, informing syntactictheoretical debates and elucidating models of bilingual language processing.
Address for correspondence

Department of Spanish, Italian, and Portuguese, 352 North Burrowes Building, The Pennsylvania State University,
University Park, PA 16802 -6203, U.S.A.; e-mail: <ajt5@psu.edu> .
The International Journal of Bilingualism

404

A. J. Toribio

2 Preliminary overview
As a convenient point of departure, the discussion commences with a cursory overview of
the language contact phenomenon of interest, and subsequently turns to survey the methodologies that subserve recent investigations into codeswitching. In so doing, the ensuing
paragraphs make explicit the rationale and methodology for the investigation that is the
kernel of the work.
2.1
Codeswitching as rule-governed bilingual behavior

Codeswitching refers to the ability on the part of bilinguals to alternate between their linguistic
codes in the same conversational event.1 With respect to its linguistic form, codeswitching
in intraturn utterances may be intersentential or intrasentential, as exemplified in the SpanishEnglish sentences in (1a-b) respectively. (For ease of exposition, the Spanish-language forms
in codeswitched examples are rendered in italics. Slashes indicate language switches in the
translations.)
(1)a. rase una vez una linda princesita blanca como la nieve. Her stepmother, the queen,
had a magic mirror on the wall.
Once upon a time there was a beautiful princess as white as the snow./Her stepmother,
the queen, had a magic mirror on the wall.
b. Por la noche, los siete enanitos found her on the ground, seemingly dead.
At night, the seven dwards/ found her on the ground, seemingly dead.
The status of intrasentential codeswitching had been much disputed in the early literature. Some linguists viewed it as indicative of imperfect language acquisition, extreme
cross-linguistic interference, or language erosion (but cf. Toribio, 2000d, 2000f), and numerous
others despaired of finding any constraints on what Lance (1975) called a willy-nilly
combination of language forms. However, subsequent studies have revealed that codeswitching is rule-governed and systematic (cf., Aguirre, 1977; Gingrs, 1974; Pfaff, 1979;
Timm, 1975), demonstrating grammatical regularities that reflect the operation of underlying syntactic restrictions (cf., Lipski, 1985; McClure, 1981; Poplack, 1980; Zentella, 1981).
For example, Spanish-English bilingual speakers will agree that the sentences in (2) represent possible codeswitches, whereas those in (3) do not, although they may be unable to
articulate exactly what accounts for this differential judgment.2
(2)a. Al cumplir ella los veinte aos, el rey invit many neighboring princes to a party.
On her 20th birthday, the king invited/many neighboring princes to a party.
1

The term codeswitching was first employed to refer to the coexistence of more than one structural system in the
speech of one individual by Jakobson, Fant, and Hall (1952), who use code in the abstract information theoretical
sense. In later writings, code has come to be synonymous with language or speech variety. For a brief overview
of codeswitching, consult Gumperz and Toribio (1999); see also the highly instructive chapters in Romaine (1995) and
the edited anthologies of Jacobson (1990), Milroy and Muysken (1995), and Auer (1998) for in-depth cross-disciplinary studies.

In the notation common to generative linguistic research, an asterisk designates an infelicitous sentence.

The International Journal of Bilingualism

Accessing bilingual codeswitching competence

405

b. Since she was unmarried, he wanted her to choose un buen esposo.


Since she was unmarried, he wanted her to choose/a good husband.
c. Princess Grace was sweet y cariosa con todos.
Princess Grace was sweet/and affectionate with everyone.
d. Juro por Dios que te casar con el primer hombre that enters this room!
I swear by God that I will marry you with the first man/ that enters this room!
e. At that exact moment, a beggar arrived en el palacio.
At that exact moment, a beggar arrived/in the palace.
(3)a. *Very envious and evil, the reina mand a un criado que matara a la princesa.
Very envious and evil, the/queen sent a houseboy to kill the princess.
b. *Out of compassion the houseboy abandonedla en el bosque.
Out of compassion the houseboy abandoned/her in the forest.
c. *La reina le ofreci a Blancanieves una manzana que haba laced with poison.
The queen offered Snow White an apple that she had/laced with poison.
d. *En la cabina vivan siete enanitos que returned to find Snow White asleep.
In the cabin there lived seven dwarfs that/returned to find Snow White asleep.
e. *Los enanitos intentaron pero no succeeded in awakening Snow White.
The dwarfs tried but did not/succeed in awakening Snow White.
Moreover, speakers furnish these judgments in the absence of overt instruction bilinguals are not instructed in how to codeswitch. And yet, just as monolingual native speakers
of Spanish and English have an intuitive sense of linguistic well-formedness in their languages,
Spanish-English bilinguals are able to rely on unconscious linguistic knowledge in distinguishing between permissible and unacceptable codeswitched forms.
As expressed by Bhatia and Ritchie (1996, p. 645), the challenge in contemporary
research on codeswitching is not whether or not it is subject to grammatical constraints but
how best to capture these constraints and how to make deeper claims about human language
in general and bilinguals mixing competence and their language acquisition in particular.
Accordingly, recent years have witnessed considerable effort devoted to examining codeswitching
within the context of Chomskys Principles and Parameters theory (Chomsky, 1981, 1986,
1993, 1995). These studies generally evaluate the extent to which codeswitching data can
be predicted by, and in so doing support, particular linguistic constructs (cf., Belazi, Rubin,
& Toribio, 1994; Di Sciullo, Muysken, & Singh, 1986; MacSwan, 1997; Rubin & Toribio,
1995; Toribio, 2000b, 2000c; Toribio & Rubin, 1996; Woolford, 1983). Our discussion will
not be detained in the elaboration of such proposals at the moment, though we return in
Section 4 to consider, albeit briefly, how the data to be presented herein may be couched within
the generative syntactic framework. Our immediate concern is in the methodologies for the
collection and selection of data on which syntactic generalizations are based.
2.2
Accessing codeswitching competence

Works addressing the grammar of codeswitching in bilingual speech have made use of a
wide variety of methodologies, chief among these, interviews and naturalistic recordings.
The International Journal of Bilingualism

406

A. J. Toribio

Unfortunately, these approaches may be of limited value in the study of linguistic competence, as they yield data that reflect the speakers competence only indirectly, at best.
Interviews and self-reports about bilingual speech are unreliable. Bilinguals often find it
difficult to remember which language was used in any particular speech exchange (cf.,
Gumperz, 1982). Moreover, the problem of self-reporting is exacerbated in situations of
social stigma (cf., Gumperz, 1971, among numerous others), as a speaker may refrain from
switching when being observed or recorded, owing to subjective factors such as the appropriateness of codeswitching to the interview situation and the esteem in which the practice
is held (cf., Toribio, 2000a; Zentella, 1997). Recordings of naturalistic utterances are met with
a more acute criticism: The linguistic performance of a speaker, in the form of natural data,
may not be indicative of that speakers underlying linguistic knowledge. Indeed, studies of
codeswitching performance in diverse bilingual communities have revealed significant variability and yielded counterexamples to many of the constraints posited (cf., for example, Pfaff,
1976; Poplack, 1980, 1981); of course, this is to be expected, since there are likewise no
exceptionless constraints on monolingual performance (cf., Poplack, 1983). Hence, we maintain with Jacobson (1977, p. 229) that since utterances containing elements from two
languages follow specific patterns of co-occurrence and display the same rule-governed
behavior that we normally associate with unilingual code, the distinction between competence and performance is applicable to the study of codeswitching. But, given performance
data alone, a researcher might erroneously conclude that there are no constraints on the form
that Spanish-English language alternation takes.
The problem adduced here is endemic to almost all of the codeswitching research
reported to date. Especially noteworthy in this respect is the work of Mahootian and Santorini
(1996), who admit only recordings of spontaneous speech on the grounds that linguistic
theory must account for natural occurrences of the data for which it has been constructed.
This focus on natural codeswitching data is incompatible with syntactic-theoretical modes
of inquiry, since the absence of violations of deep principles in spontaneous utterances cannot
be unequivocally ascribed to a constraint that exists on the speakers grammar. In assessing
a speakers competence, syntactic studies normally test his/ her ability to judge a given
sentence as a grammatical or ungrammatical string of the language; the assumption is that
the correct response indicates that the speaker has applied the principle that licenses the
structure of the intended form. Unfortunately, studies along these lines have been relatively
uncommon in codeswitching research; this in spite of the fact that the early studies of
the 1970 s (cf., for example, Aguirre, 1977; Gingrs, 1974; Jacobson, 1977; Lipski, 1978;
Timm, 1975) already indicated that there is a linguistic competence of codeswitching. As aptly
noted by McClure (1981, p. 72),
without native speakers judgments about the grammaticality of an utterance, it is often
difficult to determine whether the utterance clearly reflects the speakers competence and
so should be included in the corpus for which rules must account or whether it has been
affected by performance factors, such as lapses of attention, and hence should be excluded from consideration.

Thus, recordings of spontaneous speech must be complemented by elicitation of


speakers beliefs about ungrammatical sentences.
The International Journal of Bilingualism

Accessing bilingual codeswitching competence

407

However, as remarked in Schtze (1996, p. xi), the problems of intuition demand a


careful examination of judgments, not as pure sources of data, but as instances of metalinguistic
performance (cf., Birdsong, 1989 for additional extended discussion on this point). Therefore,
while judgments may offer insight into competence, we must concede that they are themselves subject to performance variables, and findings based on judgment data are susceptible
to the same confounding factors that plague findings founded in production samples. Speaking
to codeswitching data in particular, it can be readily observed that a speaker or researcher may
accept or reject a codeswitched sentence or response for nonlinguistic reasons. We have already
noted the problems inherent in soliciting norms of stigmatized behavior (cf., Toribio, 2000a,
2000b; Toribio & Rubin, 1996). And caution must additionally be taken in interpreting codeswitching judgments once obtained. In this light, we should reconsider MacSwans (1997)
dismissal of items such as that in (3a) which reference the ill-formedness of switching between
a determiner and its complement. His assessment (contrary to the claims of e.g., Zentella, 1981
that the same language is normally maintained across such pairings) is founded on the informally elicited judgments of two bilinguals who report that a short pause before the codeswitch
improves such forms considerably (1997, p. 247). Clearly this fact bears directly on the issue
at handit may represent the speakers attempt to comply with the injunction against switching
at this site.3 Owing to mitigating factors such as those outlined here, these previous contradictory accounts must be acknowledged, but interpreted with caution: syntactic theorization
must rely on and reflect data which are indicative of syntactic competence. We thus coincide
in Schtzes conclusion that while grammaticality judgments are indispensable forms of data
for linguistic theorizing, they require new ways of being collected and used.4
As made evident in the preceding discussion, then, methodological issues are at the
core of current debates in the characterization of codeswitching competence. The ease with
which counterexamples to any proposed generalization are found may be attributable not only
to differences in the methods of data collection, but also to the subsequent selection of the
data for which linguistic constraints are formulated (cf., Toribio & Rubin, 1996). While these
and other challenges confronting researchers in codeswitching should not be understated
(cf., Grosjean, 1998; Toribio, 2000a, 2000b), the present work undertakes to redress at least
some of the aforementioned methodological shortcomings.

3 The present study: Methodologies


The objective of the present study is in accessing bilingual codeswitching competence, while
circumventing some of the methodological difficulties that have compromised previous
research findings. To that end, three instruments of codeswitching behavior were developed
and deployed: a reading task, a recounting task, and a writing task, each described in (4). The
election of fairy tale narratives as a methodological tool for these tasks is well-motivated, as
such texts present familiar macrostructures. Of interest would be the patterns of language
3

Judgments similarly solicited from these two informants lead MacSwan to dismiss items such as those in (3d) which
reference the ill-formedness of switching after a complementizer as erroneous data and conclude that there is no
ban on switches at this juncture (1997, p. 241), contrary to the claims of for example, Gumperz (1976).

For thorough discussion of the role and use of grammaticality judgments in linguistic theory, consult Schtze (1996).
The International Journal of Bilingualism

408

A. J. Toribio

alternation that are sanctioned in narratives, and the potential differential status of codeswitched
forms across the three conditions.
(4)

Tasks

a. Reading Task
Participants are instructed to read aloud two fairy tale fragments Snow White and
the Seven Dwarfs /Blancanieves y los Siete Enanitos, which includes grammatically
unacceptable codeswitching, and The Beggar Prince/El Prncipe Pordiosero, prepared
in well-formed codeswitched sentences and respond to questions that reference readability, comprehension, enjoyability, and grammatical form. The reading recital and
responses to the questions that follow are recorded and subsequently transcribed.
b. Recounting Task
Participants are instructed to recount the ending of Snow White and the Seven Dwarfs /
Blancanieves y los Siete Enanitos or The Beggar Prince/El Prncipe Pordiosero in
codeswitching; the narratives are recorded and subsequently transcribed.
c. Writing Task
Participants are instructed to retell, in writing, the tale of Little Red Riding Hood/La
Caperucita Roja, as depicted in a sequence of color drawings; the narrations are typed
from the handwritten originals, faithfully reproducing the content and form.
These tasks were designed to combine characteristics of controlled and naturalistic
language activities. The reading task required the bilingual participants to draw on their
languages automatically, without forethought, while eliciting unconscious reactions and
explicit judgments about acceptable and unacceptable language combinations; the recounting
task was intended to engage the participants in bilingual speech production, offering a
measure of codeswitching performance via a common monological narrative activity; and
the writing task was devised to elicit texts that would be illustrative of the creativity of
bilingual code-alternation, while at once revealing of the notions of grammatical well-formedness that modulate bilingual speakers codeswitching expression. Of course, as these tasks
represent the elicitation of codeswitching behavior, the linguistic forms obtained should not
necessarily duplicate the forms observed in spontaneous speech. In this sense, the elicitation
situation before the participants is artificial the subject is being asked to demonstrate a
behavior that may be very different from his/her everyday speech mode raising the standard issues of ecological validity in linguistic research (cf., Schtze, 1996). Nevertheless, the
language samples yielded by means of all of these tasks were assumed to provide important
insights into speakers sensitivity to codeswitching norms. More generally, the methodologies would advance the aim of compiling a valid data set, establishing the constraints that
characterize Spanish-English codeswitching competence.
The objectives are accomplished by reference to the linguistic behavior of 10 speakers
Yanira, Federico, Guadalupe, Carlos, Carmen, Belinda, Emma, Sara, Noem, and Lorenzo on
the three measures. These participants were randomly selected from a larger study of SpanishEnglish bilingualism.5,6 All were native Spanish speakers of Mexican heritage who had lived
in Santa Barbara County for a minimum of 15 years at the time of observation. They were
individually tested in the language center on a university campus. Each sat at a separate
cubicle furbished with a tape-player and headset (earphones with attached microphone),
The International Journal of Bilingualism

Accessing bilingual codeswitching competence

409

contributing to the desired privacy and silence. Concise written instructions (in English or
Spanish, as requested) were presented for each task, and the sessions were untimed.7

4 Results
The ensuing discussion further expounds on the tasks and presents the data elicited. The
section ends with an analysis and synthesis of the methodologies and results.
4.1
Reading task

In the first of the codeswitching narrative tasks, participants were instructed to read two
fairy tales aloud and then respond to the questions that followed. The two narrative texts,
presented in randomized order, were of similar length and incorporated a comparable number
of switches, though they differed significantly in the type of codeswitching represented:
The Beggar Prince included switches at those boundaries that are thought to serve as
common switch sites in bilingual speech (e.g., between subject and predicate, between verb
and object, between noun and subordinate clause), and Snow White and the Seven Dwarfs
included switching at boundaries known to violate codeswitching norms (e.g., between auxiliary and main verb, between object pronoun and main or auxiliary verb, between noun and
modifying adjective). Excerpts of each fairy tale appear in (5).8
5

The study on which this work is based was carried out in 1997 98 at the University of California, Santa
Barbara, in the context of a research group convened and directed by the author. The group was motivated
by a broad interest in Spanish in the United States, and sought to explore questions pertaining to the historical and continued presence of Spanish and Spanish-English bilingualism in present-day cosmopolitan
societies, with special attention focused on the City of Santa Barbara. The aim of the study was in identifying those factors, including linguistic, social, and psychological, which influence the form of the local
language. The author gratefully acknowledges the support of the National Endowment for the Humanities
and various intramural funding agencies, among these, the Interdisciplinary Humanities Center, UC Mexus,
and the Academic Senate, and expresses sincere appreciation to seven student researchers for their commitment and effort in completing the project: Rene Basile, Mimi Beller, Cecilia Montes-Alcal, Silvia
Prez-Lpez, Christina Piranio, Guillermo Vsquez, and Patxi Zabaleta.

A review of responses by all 50 participants in the present context would entail the management of a wealth
of data, and require a concision in analysis or a synthesis of isolated linguistic features that would undermine the investigation. As the student collaborators were interested in diverse aspects of bilingual speech
(e.g., attrition and innovation of morphosyntactic structures, written vs. oral codeswitching, etc.), we sought
out for transcription and analysis the language samples of informants who produced large quantities of
speech. Within those, however, subjects were not specially chosen for the linguistic behavior (although one
informant was excluded here for her sociolinguistic attitudes owing to her negative view of codeswitching
behavior, she did not produce any intrasentential switches).

This is but one of three components of the larger study referenced above. Also administered was an extensive sociolinguistic survey; this questionnaire was developed in 1994 95 in the context of a research focus
group codirected with H.S. Gopal and Kimberly Noels. A second instrument tested participants knowledge
of diverse morphosyntactic properties of Spanish (cf., Zabaleta, 2000). All test instruments were prepared
in English and Spanish to maximize participants comfort. (A fourth narrative task is omitted from consideration here; this picture-telling task was intended to provide a base measure of speakers Spanish language
abilities (cf., Toribio, 2000a).)

Note that it is not possible to assign a base/ matrix and embedded language to these codeswitched texts; for
discussion of such notions and distinctions, see Joshi (1981), Nishimura (1986), and Myers-Scotton (1993),
among numerous others.
The International Journal of Bilingualism

410

A. J. Toribio

Narrative reading texts


a. The Beggar Prince/ El Prncipe Pordiosero

(5)

El rey Arnulfo tena una hija muy hermosa que se llamaba Graciela. Al cumplir ella los
veinte aos, el rey invit many neighboring princes to a party. Since she was unmarried,
he wanted her to choose un buen esposo. Princess Grace was sweet y cariosa con todos.
Tena solamente un defecto: she was indecisive. Surrounded by twelve suitors, she could
not decide and the king se enoj. Grit, Juro por Dios que te casar con el primer hombre
that enters this room! At that exact moment, a beggar, who had evaded a los porteros,
entr en la sala. Exclam, Acabo de or lo que dijo usted! Jur por Dios! The princess
is mine! There was no going back on such a solemn oath y el pordiosero se prepar
para la boda. Everyone was surprised to see lo bien que se vea in his borrowed clothes.
Despus de algunas semanas, the beggar made an announcement to the princess. El
nuevo esposo le dijo a la princesa that the time had come to leave the palace. They had
to return to his meager work and a house que era muy humilde
King Arnold had a very beautiful daughter named Graciela. On her 20th birthday, the
king invited/many neighboring princes to a party. Since she was unmarried, he wanted
her to choose/a good husband./Princess Grace was sweet/and affectionate with everyone.
She had only one defect:/she was indecisive. Surrounded by twelve suitors, she could
not decide and the king/became angry. He cried, I swear by God that I will marry you
with the first man/ that enters this room! At that exact moment, a beggar, who had
evaded/the doormen, entered into the room. He exclaimed, I heard what you said! You
swore by God!/ The princess is mine! There was no going back on such a solemn
oath/and the beggar prepared for the wedding./Everyone was surprised to see/how well
he looked in his borrowed clothes./After a few weeks,/the beggar made an announcement to the princess./The new husband told the princess/that the time had come to leave
the palace. They had to return to his meager work and a house/that was very humble
b.Snow White and the Seven Dwarfs/ Blancanieves y los Siete Enanitos
rase una vez una linda princesita blanca como la nieve. Su madrastra, la reina, tena
un mgico mirror on the wall. The queen often asked, Who is the ms hermosa del
valle? Y un da el mirror answered, Snow White is the fairest one of all! Very envious
and evil, the reina mand a un criado que matara a la princesa. El criado la llev al
bosque y out of compassion abandoned la all. A squirrel took pity on the princess and
led her to a pequea cabina en el monte. En la cabina, vivan siete enanitos que returned
to find Snow White asleep in their beds. Back at the palace, the stepmother again asked
the espejo: Y ahora, quin es la ms bella? El espejo otra vez le answered, without
hesitation, Snow White! The queen was very angry and set out to find the casita de
los enanitos. Disfrazada de vieja, la reina le ofreci a Blancanieves una manzana que
haba laced with poison. When Snow White bit into the apple, she call desvanecida al
suelo. Por la noche, los enanitos la found, seemingly dead
There once was a beautiful princess as white as the snow. Her stepmother, the queen, had
a magic/mirror on the wall. The queen often asked, Who is the/most fair in the valley?
And one day the/mirror answered, Snow White is the fairest one of all! Very envious
and evil, the/queen sent a houseboy to kill the princess. The houseboy took her to the
forest and/out of compassion abandoned/her there./A squirrel took pity on the princess
The International Journal of Bilingualism

Accessing bilingual codeswitching competence

411

and led her to a/small cabin in the forest. In the cabin, there lived seven dwarfs that/returned
to find Snow White asleep in their beds. Back at the palace, the stepmother again asked
the/mirror: And now who is the most beautiful? the mirror again answered/her, without
hesitation, Snow White! The queen was very angry and set out to find the/house of
the dwarfs. Disguised as an old lady, the queen offered Snow White an apple that she
had/laced with poison. When Snow White bit into the apple, she/fell fainting to the
floor. At night, the dwarfs found/her seemingly dead
By their performance, as by their assertions, sampled below, all 10 participants read the
well-formed codeswitched text, The Beggar Prince, with little effort, but had consistent problems with the ill-formed codeswitched Snow White and the Seven Dwarfs, demonstrating
various types of disfluency, including pauses, false starts, breakdowns, even laughter. Some
participants unknowingly corrected ill-formed switches in their reading, for example, by
changing she call to se call, and el mirror to the mirror; other attempts at selfcorrections included the rendering of found la as la found her. And some stammered in
producing phrases such as the the the espejo, as if ensuring that a switch was intended
at a particular inopportune juncture.9
Participants actions, however inadvertent, were substantiated by their introspections
on the two texts. As reported in (6), The Beggar Prince was judged to be easily read and
understood. Several participants believed their reading fluency owed to their facility with English
and Spanish, others reported their success due to the fact that the text reflected their own codeswitching practice. In contrast, Snow White and the Seven Dwarfs was deemed confusing,
despite participants acquaintance with the story. Some found the text unnatural and harsh,
and several offered up ways of editing the language switching to make it sound right.
(6)

Narrative reading task


Was the segment of the fairy tale easily read? Was it easily understood?

a. Comments referencing ease of readability and comprehension of The Beggar Prince


Carlos:

The segment was interesting; it was easily read and understood.

Federico: S, siento que el fragmento del cuento fue fcil de leer y fcil de entender;
porque puedo leer en los dos idiomas, me imagino que al mismo nivel; no
me caus ninguna angustia leer este cuento.
Yes, I feel that the fragment of the story was easy to read and easy to understand; because I can read both languages, I imagine that at the same level; it
did not cause me any anguish to read the story.
Sara:

No fue difcil; estoy impuesta a cambiar


It was not difficult; I am accustomed to switching

Lorenzo: I think this one flowed a little bit better; it was easier to go from back to forth
in English and Spanish; [ ] it was pretty well understood; there was no
harsh grammatical errors that made it hard to transition. [sic]
9

One reviewer suggests that a statistical analysis is warranted for these readily quantifiable miscues. However, the
paper is not grounded in psycholinguistics or applied linguistics; it does not aim to present quantifiable data, but
to attend instead to the description of language alternation and the demonstration of its rule-governed nature.
The International Journal of Bilingualism

412

A. J. Toribio

b. Comments referencing ease of readability and comprehension of Snow White and the
Seven Dwarfs
Yanira:

Too much switching made it confusing.

Carmen:

It was harder to read , and, because it was so hard to read, it was harder to
understand.

Belinda:

It was hard to shift from English to Spanish or vice versa.

Federico: Este fragmento del cuento de Blancanieves fue un poco ms difcil de leer,
no fue difcil de entender, pero se me hizo un poco ms difcil la lectura
en el aspecto de que no llevaba un ritmo, o sea, que el ritmo de la lectura
fue un poco interrumpida por el hecho que unas palabras las usaron en el cuento
en una manera que yo no las uso generalmente en ocasiones que he mezclado
el lenguaje.
This fragment of the story of Snow White was a little more difficult to
read, it was not difficult to understand, but the reading was a little more
diffcult for me in the sense that it did not have a rhythm, that is, that the
rhythm of the reading was a little interrupted by the fact that some words were
used in the story in a way that I dont generally use them on occasions when
I have mixed the language.
Sara:
haba algunas oraciones que didnt make sense
there were sentences that didnt make sense
Lorenzo: The segment of the fairy tale was somewhat easily read, although what it is
is that some of the sentences couldve changed from Spanish to English in a
better way; there are certain places that really werent really right to break
from English to Spanish or from Spanish to English. The story was easily understood because I understand English and Spanish, but I just think, like, for
example the last sentence, When Snow White bit into the apple, she call
desvanecida al suelo, that I wouldnt say it, it doesnt sound right. I would
probably say, When White bit into the apple, ella se call al suelo. Or she
fell desvanecida al suelo
The participants were then asked to compare the two texts, again on measures of readability, comprehension, and enjoyability. Consistent with their reading and evaluations of
the individual fragments, most expressed a preference for The Beggar Prince, as articulated
in (7a). There were exceptions: one participant, Lorenzo, stated that he just did not like the
stories (7b), and two other participants indicated a preference for Snow White and the Seven
Dwarfs, though, as explained in (7c), they favored the text for the well-known plot and
vocabulary, rather than for its grammatical form.
(7)

Narrative reading task


In comparing the two texts, which one was more easily read? More easily understood?
Which one did you enjoy best?

a. Comments indicating preference for The Beggar Prince


Yanira:

The Beggar Prince flowed better. You didnt get stuck on the switches
it didnt mix the languages so often.

The International Journal of Bilingualism

Accessing bilingual codeswitching competence

413

Guadalupe:I enjoyed this one, The Beggar Prince. I dont know why.
Carmen:

The first one. Why? Because it was easier to read and I actually understood
the story.

b. Comments indicating no preference


Lorenzo: Id have to say that theyre both the same; me dio igual los dos. I dont know,
I guess I really dont like stories.10
Id have to say that theyre both the same;/it was the same to me./
c. Comments indicating preference for Snow White and the Seven Dwarfs
Federico: Se me afigura que el fragmento de Caperucita Roja [ El Prncipe Pordiosero]
fue un poco ms fcil y de entenderse tambin. Me gust ms el de
Blancanieves, pero eso es porque me gusta ms ese cuento no necesariamente la manera en que est escrito, pero si tuviera yo que leerle el cuento
a otra persona me gustara leerle mejor de Caperucita Roja [ El Prncipe
Pordiosero].11
I figure that the fragment of Little Red Riding Hood [The Beggar Prince]
was a little more easy and to read too. I like the Snow White one more,
but that is because I like that story more not necessarily the way it is written,
but if I had to read the story to someone else I would rather read Little Read
Riding Hood [ The Beggar Prince].
Emma:

I think the Snow White was more easy to read, because there was some words
in The Beggar Prince that I didnt really know before so, I enjoyed the
one about Snow White and the seven dwarfs more.

Finally, participants were asked to reflect and comment specifically on the codeswitching forms represented in the two texts. All 10 participants recognized the differentiating
codeswitching patterns, which they perceived to be more abrupt, more frequent, and less
patterned in Snow White and the Seven Dwarfs than in The Beggar Prince. Sample
comments are transcribed in (8):
(8) Comments referencing codeswitching across the two texts
Yanira:

There is mixing in The Beggar Prince, but it makes sense. Snow White
changes without a pattern.

Federico: Como mencion anteriormente la diferencia el tipo de mezcla es un poco ms


inadecuada de mi punto de vista el de Blancanieves. Se me hizo un poco
ms difcil la manera en que se fragmentaron las frases del espaol al ingls.
As I mentioned previously the difference in the type of mixing is a little bit
more inadequate in Snow White in my point of view. The manner in which
10

Lorenzos expressed dislike of the stories may be attributed not to his aversion to the linguistic or grammatical form of
the stories, but to their simplicity; he produced the most creative and lengthy Beggar Prince narrative in the study.

11

Federico later corrected his error in misidentifying the fairy tale, saying, Quiero hacer una correccin a lo que
dije anteriormente. Me equivoqu con el ttulo del cuento que haba ledo. Se llama El Prncipe Pordiosero, no
Caperucita Roja; estaba confundido. I want to make a correction to what I said previously. I made a mistake in
the title of the story I had read. Its called The Beggar Prince, not Little Red Riding Hood; I was confused.
The International Journal of Bilingualism

414

A. J. Toribio

the sentences were fragmented from Spanish to English was a little more
difficult for me.
Guadalupe: I dont know, for some reason I liked The Beggar Prince. It read more
smoothly, I think.
Carmen:

I dont know really what the difference is, but, the other one [The Beggar
Prince] was half in Spanish and half in English, and so was this one [ Snow
White], but the other one was just easier to read, I dont know exactly if its
the way part of the sentence or which words you use Spanish and which you
dont.

Belinda:

The changes in Snow White were harder to understand.

Emma:

There is more a mixing in the first one, Snow White and the Seven Dwarfs

Sara:

Snow White [ ] thats not how I mix languages.

Lorenzo: The Beggar Princess didnt have such breaks in between sentences, it
didnt go where they wouldnt connect. Blancanieves if it began in English
and went into Spanish, it was a point where it shouldnt, or it just didnt
sound right.
Thus, the reading task proved useful in accessing intuitions and judgments on distinct
codeswitching forms. Specifically, all of the participants demonstrated fewer errors in
producing the language forms in The Beggar Prince, and admitted a more positive disposition towards this well-formed text, with respect to readability, comprehension, enjoyability,
and patterns of language alternation. Taken together, the participants responses revealed a
marked sensitivity to specific codeswitching patterns. However, it was deemed important to
include less controlled measures that would elicit more naturalistic bilingual behavior, and
accordingly, two storytelling tasks were administered.
4.2
Recounting task

In the recounting component, participants were instructed to select one of the fairy tale fragments previously presented and recount the ending in Spanish-English codeswitching.12 The
productions were recorded, and subsequently transcribed and analyzed for linguistic content.
All but one of the 10 story-telling narratives produced in this condition were well-elaborated
in codeswitched speech; a representative excerpt appears in (9).13 Even a cursory overview
of the oral narratives reveals a broad use of both languages, bringing into question the assumption that one language must be the base or matrix language in codeswitched speech.
(9)

Por la noche los enanitos they found uhhh Blancanieves seemingly dead. Se pusieron
muy tristes y a llorar and then one of them had an idea to bury her. Arriba en la

12

While it is unusual to divorce codeswitching production from its social context, such isolated tasks prove a necessary step in controlling for the variables that would otherwise confound the inferences drawn from the study. For
instance, an extensive background questionnaire indicated that some participants seldom engaged in codeswitching
in their natural speech productions, for lack of opportunity or inclination, and thus codeswitching had to be elicited
(cf., Toribio, 2000a).

13

Pauses or breaks in the narration are marked with ellipses ( ).

The International Journal of Bilingualism

Accessing bilingual codeswitching competence

415

montaa donde estuviera rodeada por todos sus heridos queridos all the little birds,
the little possums, all the little animals of the forest because she loved them so much
Entonces se la llevaron, este in a procession they marched up there. Y como, como
eran muy imaginativos ellos, muy este they, they built a casket of ahhh clear
crystal casket. Y all es donde la metieron y la velaron por un da, dos das, y todos los
animalitos del bosque were there with them all sad because she was a very beautiful
doncella Ya que haban pasado unos dos tres das pas por all a very handsome prince era alto, moreno de ojos grandes, nice long lashes he was just
passing by because andaba buscando a su amor perdido y de repente he sees
Blancanieves in the clear casket y l saba, algo le dijo en su corazn que ella era
ella era la persona soada, la persona que andaba buscando toda su vida y su
belleza took him by surprise he wanted to see her up close y se le acerc y la sac
del atad and without knowing why he kissed her on the lips. En eso Blancanieves
despert de un sueo tan profundo el prncipe la haba sacado del abismo As she
opened her eyes she saw the most handsome beautiful prince y entonces Blancanieves
supo que tambin era el amor de su vida y se fueron a vivir una vida hermosa,
llena de amor y pues, colorn, colorado este cuento se ha acabado. (Sara)
At night the dwarfs / they found uhhh Snow White seemingly dead.14/ They became
very sad and began to cry /and then one of them had an idea to bury her./Atop the
mountain where she would be surrounded by her wounded loved ones /and the little
birds, the little possums, all the little animals of the forest because she loved them so
much / Then they took her, uhhh in a procession they marched up there./And as, as
they were very imaginative, very uhhh / they, they built a casket of ahhh clear
crystal casket./And there is where they put her and they mourned her a day, two days,
and all of the animals of the forest/ were there with them all sad because she was a
very beautiful/ maid When there had passed some two three days there passed
by there/a very handsome prince / he was tall, dark with big eyes,/nice long lashes
he was just passing by because / he was looking for his preferred love and
suddenly/he sees Snow White in the clear casket/and he knew, something told him in
his heart that she was she was the dream person, the person that he had been searching
for all of his life and her beauty/took him by surprise he wanted to see her up
close/and (he) got nearer to her and he took her out of the casket/and without knowing
why he kissed her on the lips./ At once Snow White awoke from a deep sleep the
prince had brought her out of an abyss /As she opened her eyes she saw the most handsome beautiful prince /and then Snow White knew that he too was the love of her
life and they left to live a splendid life, filled with love and well, thats all folks.
For all participants in this condition, the vast majority of language switches occurred
at sentence boundaries, many preceded by pauses signaling principal discourse breaks required
in recalling and reformulating the story. The narratives additionally included other stylistic
features commonly marked by language alternations in bilingual speech; as outlined in (10),
some of these stylistic strategies are especially germane to storytelling (cf., Gumperz,
1976, 1982; Montes-Alcal, 2000; Valds, 1976; Zentella, 1981, 1997).
14

Switch boundaries that border on the proper names of fairy tale characters have generally been excluded from analysis,
as names could be more salient in one or the other language.
The International Journal of Bilingualism

416

A. J. Toribio

(10) Stylistic language alternations


a. Switching for reported speech
Lorenzo: pero dijo ella, Ill give it some time,
but she said,/ Ill give it some time,
b. Switching for repetition or emphasis
Yanira:
un prncipe, Prince Charming estaba pasando por el bosque
a prince,/ Prince Charming / was passing through the forest
Belinda: un gran palacio, a great palace, y all entonces la princesa
a great palace,/a great palace,/and there the princess then
c. Switching for qualification or elaboration
Sara:
por todos sus heridos queridos all the little birds, the little possums, all
the little animals of the forest
for all of her wounded loved ones all the little birds, the little possums, all
the little animals of the forest
Lorenzo: qu tipo de animales haban, what type of trees, flowers
what kinds of animals there were,/ what type of trees, flowers
Lorenzo: she wanted to experiment, quera ver qu haba all fuera
she wanted to experiment,/she wanted to see what was out there
Lorenzo: No haba cuartos, there was no living room, there was no, not even a bathroom.
There were no rooms,/there was no living room, there was no, not even a bathroom.
d. Switching for fixed or formulaic phrases
Belinda: Y as vivieron, they lived happily ever after.
And they lived that way,/ they lived happily ever after.
Belinda: she met this lobo feroz that asked where she was going.
she met this/ fierce wolf / that asked where she was going.
Also attested in participants oral productions were lexical insertions and tag-switches.
Lexical insertions, exemplified in (11a), represent the introduction of individual items into
a recipient language, as occasioned by unavailability or temporary lapses in memory; these
insertions often trigger a language switch for ensuing material. Tag-switches, such as okay,
so, pues well, and verdad right function as sentence fillers or reveal a speakers disposition towards the content of an utterance; they typically occur at phrase or clause boundaries,
as in the Example (11b).15
(11) Other features common in bilingual speech
a. lexical insertions
Sara:

15

because she was a very beautiful doncella Ya que haban pasado unos
das
because she was a very beautiful/ maidWhen there had passed some days

As expected, lexical insertions and tags may be evidenced in both monolingual and bilingual modes of interaction; in
contrast, codeswitching, of interest here, is illustrative of a bilingual speech mode which requires a high degree of bilingual
competence.

The International Journal of Bilingualism

Accessing bilingual codeswitching competence

Belinda:

417

ella estaba acostumbrada a todas las, umm, luxuries of her palace


she was accustomed to all of the,/ ummm, luxuries of her palace

b. tag-switches
Lorenzo: se qued unos, you know, ella dijo, Me voy a quedar aqu un mes
she stayed some days,/you know,/she said, Im going to stay here a month
Although intersentential switches predominated in the oral narratives, there were also
attested numerous examples of intrasentential codeswitching, especially at major phrase
boundaries. The excerpts shown in (12) illustrate switching between clauses (12a), between
coordinated clausal conjuncts (12b), between coordinated conjuncts (12c), between subject
and predicate (12d), between verb and complements (12e), between noun and relative clauses
(12f), and between clause and sentential modifiers (12g).
(12) Codeswitching produced in narrative story-telling task
a. Between sentential clauses, with pause
They dont know what to do and they pick her up, y la llevan a la casa
They dont know what to do and they pick her up,/and they take her to the
house
Guadalupe: They prepared for a funeral, y pusieron muchas flores
They prepared for a funeral,/and they put many flowers
Emma:
He saw that she was very beautiful, y la bes.
He saw that she was very beautiful,/and he kissed her.
Se pusieron muy tristes y a llorarand then one of them had an idea to bury her.
Sara:
They became very sad and began to cry /and then one of them had an idea
to bury her.
Yanira:

b. Between coordinated clausal conjuncts16


Se asom a la casa de los enanitos and he saw that
Yanira:
He got closer to the dwarfs house/and he saw that
Lleg un prncipe y vi a Blancanievesand he approached her and gave her a kiss.
Noem:
The prince arrived and saw Snow White/and he approached her and gave her
a kiss.
Sara:
y la sac del atad and without knowing why he kissed her on the lips.
and he took her out of the casket/and without knowing why he kissed
her on the lips.
c. Between coordinated conjuncts
Carlos:
Her mother le habl and sent her to make, to take
Her mother/spoke to her/and sent her to make, to take
Sara:
He wanted to see her up close y se le acerc.
He wanted to see her up close /and (he) got nearer to her.
16

The possibility of null subjects in Spanish makes it difficult to distinguish between coordination of full clauses and
coordination of predicates; the analysis here errs on the side of conservatism: coordination of clauses must include
two distinct subjects, as indicated by overt content or by verbal morphology.
The International Journal of Bilingualism

418

A. J. Toribio

Lorenzo: accepted the decision that the husband had made y se fueron al bosque
accepted the decision that the husband had made/and they went to the
forest
d. Between subject and predicate
Ellos se enamoraron y el prncipe wanted to get married.
Emma:
They feel in love and the prince/wanted to get married.
Y todos los animalitos del bosque were there with them.
Sara:
And all of the little animals of the forest / were there with them.
Sara:
pas por all a very handsome prince
there passed by there/ a very handsome prince
Sara:
y su belleza took him by surprise.
and her beauty/ took him by surprise.
Noem: Y Blancanieves y el prncipe y los siete enanitos were very happy.
And Snow White and the prince and the seven dwarfs /were very happy.
e. Between verb and complements
Carmen: Al fin, ella decidi to go back to her palace.
In the end, she decided/ to go back to her palace.
Blancanieves despert y los enanitos estaban very happy.
Noem:
Snow White awoke and the seven dwarfs were / very happy.
Lorenzo: She wanted to see la belleza que tena
She wanted to see/ the beauty that it had
Lorenzo: Nunca haba salido out of the forest before.
She had never gone/out of the forest before.
f. Between noun and relative clause
Lorenzo: se fueron al bosque where they would both live in the little house.
They went to the forest/ where they would both live in the little house.
Lorenzo: the beauty que el bosque le daba.
the beauty/ that the forest offered her.
g. For sentential modifiers
Carlos:
While on her way, se top con el lobo
While on her way,/she ran into the wolf
Al mismo tiempo, the wolf continued on the original path
Carlos:
At the same time,/the wolf continued on the original path
Carlos:
When Caperucita arrived to her grandmothers le pregunt por qu tena
dientes tan grandes.
When/Little Red Riding Hood/arrived to her grandmothers/she asked her why
she had such big teeth.
Lorenzo: Al llegar a la casita, she noted
On arriving at the little house,/she noted
Lorenzo: durante una semana, it started getting to her
within one week,/ it started getting to her
The International Journal of Bilingualism

Accessing bilingual codeswitching competence

419

While the grammatically-sanctioned intrasentential switches illustrated in (12) proceeded


smoothly, others, representing potential ill-formed intrasentential switches, gave rise to
disfluencies. For example, as shown in (13), switching after a coordinating conjunction,
subordinating complementizer, and determiner is preceded by pauses or prevented by an
immediate reiteration.17 Moreover, there were no incidences of codeswitching at the boundary
between auxiliary and main verb, or between negative marker and verb, or between demonstrative and noun, among other syntactic junctures where grammatical norms do not favor
codeswitching (recall the discussion in Section 1). In fact, not one of the participants who
elected to tell the ending to the Snow White and the Seven Dwarfs fairy tale replicated the
ill-formed switching modeled in the exemplar; this finding is significant, as it suggests that
the codeswitching in the fairy tale model was indeed incompatible with proficient bilingual
production.18
(13) Circumventing ill-formed switches
a. Disfluency
Yanira:

not frowns but un poco tristes


not frowns but /a little sad

Yanira:

they continued working and en eso iba pasando un prncipe


they continued working and/at that time there was passing a prince

Belinda:

dijo que all of a sudden


he said that /all of a sudden

Sara:

passing by because andaba buscando a su amor perdido


passing by because / he was looking for his preferred love

Noem:

pens que, Snow White was dead


he thought that,/Snow White was dead.

b. Repair
Guadalupe: el the queen was not happy that she was still alive 19
the (masc.) / the queen was not happy that she was still alive
Noem:

Ella se she came back to life


She was/she came back to life

Noem:

y umm he and he fought with her and he killed her.


and/ umm he and he fought with her and he killed her.

17

The low incidence of such switching, reported in the early work of the 1970 s, has been tested with experimental
measures by Toribio (2000b).

18

For discussion of the grammatical competence that is required for successful codeswitching, consult Toribio (2000b).
Therein, I discuss several studies of child and adult bilingual populations which reveal that increased competence in
the component languages is a prerequisite for rule-governed codeswitching (e.g., Bhatia & Ritchie, 1996; McClure,
1981; Meisel, 1989; Poplack, 1980; Rakowsky, 1989; Toribio, Roebuck, Lantolf, & Perrone, 1993).

19

This example does not represent a correction for misassignment of gender; the remainder of the oral text confirms
that reina queen is marked with the default masculine gender in the speakers lexicon. Her uncertainty regarding grammatical gender in Spanish is verified in the written codeswitching task and in a separate Spanish-language narrative
task, not reproduced here (cf., Toribio 2000a, 2000d, 2000f).
The International Journal of Bilingualism

420

A. J. Toribio

Lorenzo: con las, with the leaves of the trees.


with the,/ with the leaves of the trees.
Lorenzo: Y, the house was just a one-room house.
And,/ the house was just a one-room house.
The remaining participant, Federico, experienced difficulty in meeting the demands
of the task, and gave verbal expression to his frustration. As shown in the transcript (14), he
recounted the ending of The Beggar Prince fully in Spanish, and was apologetic in his
resignation before the task. He made a second attempt to comply with the instructions, this
time relating the segment fully in English, with a single insertion of an adverbial modifier.
(14) y ella se hizo una mujer muy humilde y muy buena de corazn. Nunca ms quiso todas
las riquezas que quera antes porque antes era una, una nia fresa, no trataba bien a
sus compaeros, y ahora era ms buena de corazn como ven se me hizo un poco difcil
mezclar el ingls con el espaol. No es que no lo puedo hacer pero casi siempre pienso
o estoy pensando en espaol. No estoy pensando en ingls y y y a veces es ms es
ms fcil terminarlo de una manera pero, y I guess I could do it both ways. I dont
know. Its hard for me though, yeah, you know if I start talking in one language I keep
talking in one language so its kind of hard. I cant concentrate on doing that. Umm so
Im not gonna try to do it umm she lived happily ever after, humildemente, umm,
without all the riches that the see, I cant, I dont know, for some reason I hold myself
back sometimes. (Federico)
and she became a very humble woman and very good at heart. She never wanted all
of the riches that she wanted before because before she had been a spoiled child, she didnt
treat her peers well, and now she was better at heart as you can see it was difficult for
me to mix Spanish with English. It isnt that I cannot do it but I almost always think or
am thinking in Spanish. I am not thinking in English and and sometimes its more
its easier to finish it one way but, and / I guess I could do it both ways. I dont know.
Its hard for me though, yeah, you know if I start talking in one language I keep talking
in one language so its kind of hard. I cant concentrate on doing that. Umm so Im
not gonna try to do it umm she lived happily ever after,/humbly,/ umm, without all
the riches that the see, I cant, I dont know, for some reason I hold myself back sometimes.
However, Federicos behavior does not impugn the validity of the storytelling task;
rather, it makes evident that not all bilinguals possess the requisite communicative competence to engage in codeswitching. As often noted in the research literature (cf., for example,
Aguirre, 1977; Valds, 1976), competence in two languages is a necessary precondition, but
insufficient prerequisite in determining successful codeswitching performance: membership in a community in which codeswitching is practiced may also be a required.20 This
mitigating factor was controlled for in the narrative writing task that ensued.
20

An extensive background questionnaire revealed that Federico does not belong to such a speech community; he participates in largely monolingual-speaking Spanish and English-language communities, so it is not surprising that a
codeswitching storytelling performance would prove arduous for him (cf., Toribio, 2000a).

The International Journal of Bilingualism

Accessing bilingual codeswitching competence

421

4.3
Writing task

In the final narrative task, participants were asked to review a sequence of color pictures depicting
the Little Red Riding Hood fairy tale and recount the story in mixed speech, but this time,
in written form. To be sure, this task was not intended to render speech samples that would
be equivalent to the oral narratives produced in the task preceding. Lipski (1985) has pointed
to the obvious limitations inherent in the use of written samples of bilingual codeswitching
as representative of speaker norms (cf., Montes-Alcal, 2000). However, the task at hand did
not elicit literary artifacts, but orthographic renditions of a renowned fairy tale narrative, and,
as such, the texts were expected to more closely approximate unmarked verbal behavior than
prose or poetry.21 Thus, while the task was designed to examine speakers creative manipulation of two languages, its greater purpose was to evoke speakers notions of codeswitching
well-formedness, while abstracting away from the demands of performance and practice.
All participants successfully completed the task, producing comprehensive fairy tale
narratives that were permeated by intrasentential switches. An excerpt of Lorenzos writing
sample is reproduced in (15), without editorial correction (save for the italics that identify
the Spanish-language segments).
(15) Once upon a time, en un lugar lejano, there was a little girl que se llamaba Caperucita
Roja. She liked walking through the forest, escuchando los pajaritos, oliendo las flores,
y apreciando la belleza natural. One day, su mam la mando a la casa de su abuelita.
She told L.R.R.H. que su abuelita ocupaba some medicine for some sickness. Caperucita
went to her grandmothers house, a llevarle medicina. On the way to the house, se topo
con un lobo feo. The wolf gave her flowers and asked where she was heading. Ella le contesto
que iba a la casa de su abuelita. The wolf then left her and headed to the grandmothers
house, corriendo lo mas pronto posible par ganarle a Caperucita. When he arrived, se
metio y persiguido a la abuelita. Caperucita luego lleg and began to knock. The wolf
changed into the grandmothers clothes y se metio de bajo de las cobijas. L.R.R.H. went
inside and noticed que su abuelita looked different. She began to ask questions about
the wolfs nose, eyes, ears, and mouth. Cuando Caperucita asked the last question, el lobo
brinc y la empeso a corretiar. At the same time, a little squirrel had warned a hunter
in the forest about this big bad wolf. El hombre corrio en seguida, a la casa de la abuelita.
There he found the wolf, persiguiendo a Caperucita. L.R.R.H. ran outside, y el lobo la
siguio. Entonces el hombre levanto su rifle and fired it at the wolf, matandolo. Caperucita
entonces empeso a llorar. The man asked her porque estaba llorando, y ella le dijo que
porque el lobo se haba comido a su abuelita. At that same time, the grandmother came
out of the dogs house. Ella se haba escondido el la casa del perro durante todo ese tiempo.
Asi que feliz termina este cuento de Little Red Riding Hood. [sic] (Lorenzo)
Once upon a time,/in a far away place,/there was a little girl/ who was named Little Red
Riding Hood./She liked walking through the forest,/listening to the little birds, smelling
the flowers, and appreciating the natural beauty./One day,/her mother sent her to her grandmothers house./She told L.R.R.H./that her grandmother needed some medicine for
21

For relevant discussion on the narrative structure of codeswitching, the reader is referred to Lipski (1985), MontesAlcal (2000), Torres (1997), Valds (1976), and the literature grounded in Keller (1979).
The International Journal of Bilingualism

422

A. J. Toribio

some sickness. Little Red Riding Hood went to her grandmothers house,/to take her
medicine./On the way to the house,/she ran into an ugly wolf./ The wolf gave her flowers
and asked where she was heading./She answered him that she was going to her grandmothers house./The wolf then left her and headed to the grandmothers house,/running
the fastest possible to beat Little Red Riding Hood./ When he arrived, he got in and
followed the grandmother. Little Red Riding Hood later arrived/and began to knock. The
wolf changed into the grandmothers clothes/and got under the covers./L.R.RH. went
inside and noticed/ that her grandmother/ looked different/She began to ask questions
about the wolfs nose, eyes, ears, and mouth./When Little Red Riding Hood/asked the
last question,/the wolf jumped and started to chase her./At that same time, a little squirrel
had warned a hunter in the forest about this big bad wolf./The man ran at once, to the
grandmothers house./There he found the wolf,/following Little Red Riding Hood./L.R.R.H.
ran outside,/and the wolf followed her. Then the man raised his rifle/and fired it at the
wolf,/ killing it. Little Red Riding Hood then began to cry./ The man asked her/ why she
was crying, and she told him that because the wolf had eaten her grandmother./At that
same time, the grandmother came out of the dogs house./She had been hiding in the dogs
house during all that time. So happily ends the story of / Little Red Riding Hood.
As in the oral production task, the participants demonstrated a broad use of both
languages in achieving a diversity of stylistic effects, for example, in signaling a change in
roles (16a), marking direct or indirect speech (16b), shifting for declarative or interrogative
(16c), attracting attention (16d), heightening an interjection or exclamation (16e), and inserting
fixed or formulaic phrases (16f).
(16) Stylistic codeswitching
a. Switching for change in roles
Sara:

Oh grandma what a big nose you have Para olerte mejor mijita [sic]
Oh grandma what a big nose you have/The better to smell you my child

b. Switching for reported speech


El wolf le pregunto where are you going Caperucita? And she told him A
Noem:
la casa de mi abuelita [sic]
The/wolf/asked /where are you going Little Red Riding Hood? And she
told him /To my grandmothers house.
c. Switching for declarative or interrogative
El lobo le pregunt where did she live
Emma:
The wolf asked her/ where did she lived
d. Switching for attracting attention
Belinda: Mira so that you get to her house sooner vete por este camino
Look/so that you get to her house sooner/take this path
e. Switching for interjections or exclamation
Sara:
Oh que bien, where does she live? [sic]
Oh good,/where does she live?
f. Switching for fixed or formulaic phrases
Carmen: Once upon a time haba una nia llamada Caperucita Roja
The International Journal of Bilingualism

Accessing bilingual codeswitching competence

423

Once upon a time/there was a girl named Little Red Riding Hood
Emma:

Y la historia termina con un happy ending.


And the story ends with a/ happy ending.

And just as the oral narratives, the written narratives incorporated lexical insertion of
culturally salient nouns such as wolf, ardillita squirrel, hunter, mom, abuelita
grandmother, grandma, and mhija my child; the example in (17) demonstrates two such
lexical insertions in one sentence.22
(17) Lexical insertion
Federico: On the way to abuelas house she runs into the lobo and he asks her where
she is going. After she tells him they both go on their way.
On the way to/grandmothers/ house she runs into the/wolf/and he asks her
where she is going. After she tells him they both go on their way.
A longer view of the Spanish and English segments of discourse produced in these
written narratives reveals that the participants generally possess an advanced degree of
bilingual competence; this ability was invoked in producing texts that incorporated intersentential
and intrasentential codeswitching. Interestingly, the intersentential alternations attested in
the written narratives are significantly reduced when compared with those produced in the
oral narratives; intrasentential codeswitching prevailed in this task, as evidenced in the examples in (18). (Note, again, that the written texts are faithfully reproduced, without editing.)
(18) Codeswitching in written narratives
a. Between coordinated clausal conjuncts

22

Yanira:

They both finish talking y el lobo camina en otra direccin


They both finish talking/and the wolf walks in another direction

Yanira:

Le dispara and the wolf dies.


He shoots/and the wolf dies.

Carmen:

Ella anda perdida so she asks him for directions


She is wandering lost/so she asks him for directions

Carmen:

El lobo llega a la casa de la abuela and he hides her in the closet


The wolf arrives at the grandmothers house/and he hides in the closet

Belinda:

She was a very nice, and sweet child, y todos la querian mucho. [sic]
She was a very nice, and sweet child,/and everyone loved her very much.

Belinda:

She thanked him for this advice y se despidieron.


She thanked him for this advice/and they took leave.

Emma:

knew of a short cut y Caperucita se fue por ahi. [sic]


knew of a short cut /and Little Red Riding Hood went that way.

But unlike the oral narrative production task, this third task produced no tag switches, as expected, given their typical
functions in oral discourse. More notably, unlike the previous condition, the written narrative task gave rise to various
types of cross-linguistic transfers. Examples such as la house of the grandmother, modeled on la casa de la abuela,
and hablar con estrangers modeled on talk to strangers, demonstrate the influence of the structure of one language
on the other.
The International Journal of Bilingualism

424

A. J. Toribio

b. Between coordinated phrasal conjuncts


Yanira:

with flowers in one hand y una canasta en el otro brazo.


with flowers in one hand/and a basket in the other arm.

En eso llega la nia saludndola but points to her big teeth,


At that moment the girl arrives greeting her/but points to her big teeth,
Yanira:
so frightened pero alegre de estar reunida con su grandma otra vez.
so frightened/but happy to be reunited with her grandmother again.
Federico: She suspects something y le hace muchas preguntas.
She suspects something/and asks her lots of questions.
Federico: Caperucita and her grandmother reunite y viven felices el resto de sus vidas.
Little Red Riding Hood and her grandmother reunite/and they live happily the
rest of their lives.
Llego a la casa de la habuela and scared her off. [sic]
Carlos:
He arrived at the grandmothers house/and scared her off.
Carmen: he hides her in the closet y le quita su gorro y se mete en la cama
he hides her in the closet/and takes off her cap and gets into the bed
Caperucita lo encontro sospechoso and pointed out that [sic]
Emma:
Little Red Riding Hood finds him suspicious/and pointed out that
Yanira:

Sara:

El hombre lobo se le acerco a Caperucita and said where are you going
mija. [sic]
The man wolf got nearer to Little Red Riding Hood/and said where are you
going my child.

Lorenzo: Caperucita luego lleg and began to knock.


Little Red Riding Hood later arrived/ and began to knock.
Lorenzo: The wolf changed into the grandmothers clothes y se metio debajo de las
cobijas. [sic]
The wolf changed into the grandmothers clothes/and got under the covers.
Lorenzo: el hombre levant su rifle and fired it at the wolf
the man raised his rifle/and fired it at the wolf
c. Between subject and predicate
Federico: The story empieza en que su mam le hace un encargo
The story/begins with her mother making a request
Emma:

Emma:
Sara:

Un dia la mam de Caperucita Roja asked Caperucita to take some food


One day Little Red Riding Hoods mother/asked Little Red Riding Hood to
take some food
When the wolf she iba a comer a Caperucita the hunter came [sic]
When the wolf/ was going to eat Little Red Riding Hood/the hunter came
cada quien went on their own way.
each one/ went on their own way.

Lorenzo: noticed que su abuelita looked different.


noticed/ that her grandmother looked different.
The International Journal of Bilingualism

Accessing bilingual codeswitching competence

425

Lorenzo: Cuando Caperucita asked the last question


When Little Red Riding Hood/ asked the last question
d. Between verb and subordinate clause
Emma:

Caperucita told him que iba a visitar a su grandmother.


Little Red Riding Hood told him/ that she was going to visit her/grandmother.

Emma:

El lobo le pregunt where did she live


The wolf asked her/ where did she live

Sara:

after telling him donde vive la grandma cada quien went on their own way.
after telling him where the grandma lived each one/ went on their own way.

Lorenzo: She told L.R.R.H. que su abuelita ocupaba


She told L.R.R.H./ that her grandmother needed
Lorenzo: noticed que su abuelita looked different.
noticed/ that her grandmother looked different.
Lorenzo: The man asked her porque estaba llorando. [sic]
The man asked her/why she was crying.
e. Between verb and complements
Yanira:

La madre de Caperucita le da a jar of honey.


Little Red Riding Hoods mother gave her/a jar of honey.

Federico: In the title frame we see Caperucita Roja en el bosque


In the little frame we see/ Little Red Riding Hood in the forest
Guadalupe: pidi que ella lleva some food to her gradmother [sic]
asked to take/some food to her grandmother
Carmen:

siempre traa a red cloth over her head.


always wore/a red cloth over her head.

Carmen:

su abuelita que estaba sick.


her grandmother who was/sick.

Emma:

The wolf asked Caperucita a donde iva. [sic]


The wolf asked/ Little Red Riding Hood where she was going.

Sara:

para que le llevara food to her grandmother.


to take/ food to her grandmother.

Lorenzo: que su abuelita ocupaba some medicine for some sickness.


that her grandmother needed/some medicine for some sickness.
Lorenzo: Caperucita went to her grandmothers house a llevarle medicina.
Little Red Riding Hood went to her grandmothers house/to take her medicine.
f. Between preposition and objects
Yanira:

El lobo platica con Little Red Riding Hood for a while.


The wolf chatted with/ Little Red RidingHood for a while.

Federico: goes to la casa de la abuelita


goes to/ the grandmothers house
The International Journal of Bilingualism

426

A. J. Toribio

Carlos:

and sent her to take honey to su abuela.


and sent her to take honey to/ her grandmother.

Sara:

recuerda no hables con strangers along the way


remember dont talk with/strangers along the way

g. Between noun and relative clause


Lorenzo: there was a little girl que se llamaba Caperucita Roja.
there was a little girl/ who was named Little Red RidingHood.
h. For phrasal modifiers
Carmen:

directions para llegar a la casa de su abuela.


directions/to get to her grandmothers house.

Emma:

Esta es la historia of the Little Red Riding Hood.


This is the story/of the Little Red Riding Hood.

i. For sentential modifiers


Yanira:

As Little Red Riding Hood is walking along the forest se encuentra con un lobo.
As Little Red Riding Hood is walking along the forest/she encounters a wolf.

Guadalupe: y llego when the wolf was chasing Little red riding Hood. [sic]
and he arrived/ when the wolf was chasing Little Red Riding Hood.
Carlos:

to take this path por que era mas corto. [sic]


to take this path/ because it was shorter.

Carlos:

At the same time Caperucita encontro a su habuela. [sic]


At the same time/ Little Red Riding Hood found her grandmother.

Carmen:

Her mom had given her some soup para que le llevara a su abuela
Her mom had given her some soup/ to take to her grandmother

Carmen:

se mete en la cama to pretend that he is the grandmother


he gets into the bed/ to pretend that he is the grandmother

Belinda:

en el camino she met this lobo feroz that asked where she was going.
on the path/she met this/fierce wolf/that asked where she was going.

Belinda:

so that you get to her house sooner vete por este camino
so that you get to her house sooner/take this path

Emma:

to her grandmothers house porque su abuelita estaba enferma.


to her grandmothers house/ because her grandmother was sick.

Lorenzo: Once upon a time, en un lugar lejano, there was a little girl
Once upon a time,/ in a place far way,/ there was a little girl
Lorenzo: On the way to the house, se topo con un lobo feo. [sic]
On the way to the house/she ran into an ugly wolf.
Lorenzo: headed to the grandmothers house, corriendo lo ms pronto posible
headed to the grandmothers house,/running the fastest possible
Thus, the written mode did not constrain but encourage switching at a diversity of
syntactic junctures. Still, though the number and variety of switch sites increased, the written
The International Journal of Bilingualism

Accessing bilingual codeswitching competence

427

narratives were similar to the oral narratives in revealing a preference for switching at major
syntactic boundaries.
4.4
Analysis and synthesis of the results

The findings for the reading, recounting, and writing components are robust across subjects
and tasks. By their reading performance, as by their introspection on the differential codeswitching in the two model texts, the participants demonstrated a sensitivity to grammatical
coherence in codeswitching patterns. While the language alternations in The Beggar Prince
were thought to be systematic and more correct, the switches in Snow White and the Seven
Dwarfs were rejected as affected and forced. Several participants involuntarily self-corrected
the ill-formed switches in their readings of this fairy tale, and others proposed explicit editing
recommendations for improving on the ill-formed combinations in the text. Since the participants would later demonstrate their ability at alternating easily and effortlessly between
languages, the repeated performance errors that they evidenced in reading particular
Spanish-English combinations could be argued to have their origin in grammatical violations.23
And the participant, Federico, who was unable to produce an oral codeswitched narrative was
nevertheless successful in identifying and articulating the differences between the texts in
the reading task. These findings are telling of the unconscious knowledge that licenses
permissible switches and disallows unacceptable alternations in bilingual speech.
The controlled reading task was complemented by two more natural activities in
which participants were required to recount fairy tales in oral and written codeswitching. The
productions in both conditions converged in reflecting a strict compliance to codeswitching
norms. The oral narratives incorporated intersentential alternations and intrasentential codeswitching at major phrasal boundaries, in addition to the lexical insertions and formulaic
expressions frequent in bilingual (and monolingual) speech. Significantly, language switches
that would violate phrasal coherence were circumvented or repaired, such that there were no
true violations of codeswitching norms attested. The preponderance of language alternations at sentence boundaries may be explained by the condition a lone speaker directing
speech into a microphone in an individual carrel which precludes the casual interchange
in which intrasentential codeswitching is favored.24 The participants were able to more fully
draw on their bilingual resources in the writing task. The quantity and types of intrasentential
codeswitching were significantly increased in the writing activity. In fact, a careful consideration of the codeswitching produced in this condition, especially in view of the reduced
codeswitching modeled for them in the The Beggar Prince fairy tale, attest to the participants skillful interpolation of Spanish and English. The abundance of intrasentential
codeswitching was likely fostered by the specifics of the task: the visual aids and lack of time
constraints for formulating the story were certain to reduce anxiety. However, the task engendered reflection and self-correction, and as a consequence, codeswitching well-formedness
23

Consult Poulisse (1999) for informative discussion of recent developments in research on slips of the tongue, briefly
noted in Section 4.2.

24

This speaks directly to the need for complementing natural elicitation techniques with grammaticality judgments:
Via elicitation of judgments, we can examine sentence types that might not be produced in spontaneous or controlled
speech situations.
The International Journal of Bilingualism

428

A. J. Toribio

was maintained. These differential and at once convergent findings speak to the need for multiple
measures in the study of codeswitching and to the reliability of the measures and the data
reported herein.

5 Significance of the findings


Having established the reliability of the elicited data, we now turn to determine how these
findings can be accommodated within generative syntactic theory, and end by presenting directions for further research in bilingual language processing.
5.1
The formalization of the attested coherence and co-occurrence constraints

As noted in Section 1, recent studies have explored codeswitching within syntactic-theoretical frameworks (Chomsky, 1981, 1986, 1993, 1995). These investigations have sought an
explanatory adequacy that was lacking in earlier, more descriptive formulations, by exploiting
universal principles and relations that are hypothesized to characterize monolingual competence (cf., Muysken, 1995). This line of inquiry into language contact was initiated by
Woolford (1983). According to Woolford, though the lexicons of the two component grammars of the bilingual remain separate, thereby precluding word-internal switching (19a), in
codeswitched speech each grammar contributes part of the sentences, that is, both lexicons
have access to terminal nodes in syntactic constructions that are common to both languages
(19b d). In contrast, whenever a phrase structure rule unique to one language is used to
expand a node, the terminal positions must be filled from the lexicon of that language,
predicting the ill-formedness of examples such as those in (20), where the phrase structures
of English and Spanish differ.25
(19) a. *I am readiendo./*(Yo) estoy leying.
I am read/ing.
b. I put the forks en las mesas. (McClure, 1977, cited in Woolford, 1983)
I put the forks /on the tables.
c. Todos los mexicanos were riled up. (Pfaff, 1979, cited in Woolford, 1983)
All of the Mexicans/ were riled up.

25

Woolfords model represents an early generative reformulation of two constraints previously proposed in the literature: Poplacks (1980) Free Morpheme Constraint and Equivalence Constraint. The Free Morpheme Constraint accounts
for the nonoccurrence of word-internal switching; and the Equivalence Constraint predicts that codeswitching will be
permitted where the grammars of Spanish and English coincide, but not where they diverge. Lipski (1978) and Pfaff
(1979) likewise conclude that surface structures common to both languages are favored sites for switching.
(i) Free Morpheme Constraint
A switch may occur at any point of the discourse at which it is possible to make a surface constituent cut and still
retain a free morpheme.
(ii) Equivalence Constraint
Codes will tend to be switched at points where the surface structures of the languages map onto each other.

The International Journal of Bilingualism

Accessing bilingual codeswitching competence

429

d. El hombre who saw the accident es cubano. (Gingrs, 1974, cited in Woolford, 1983)
The man/ who saw the accident is Cuban.
(20) a. *El man viejo est enojado. (Gingrs, 1974, cited in Woolford, 1983)
The old/man / is angry.
b. *El hombre old est enojado. (Gingrs, 1974, cited in Woolford, 1983)
The old/man / is angry.
c. *Yo lo bought. (Quintero, cited in Woolford, 1983)
I bought/it.
d. *Yo it compr. (Quintero, cited in Woolford, 1983)
I bought/ it.
While Woolfords model goes a long way towards accounting for codeswitching data
elicited in our study no participant produced forms of the type represented in (19a) or
(20) a closer examination reveals that it is insufficiently restrictive, allowing the grammar
to overgenerate. For example, as has been observed in the research literature (and corroborated in their omission in the texts reproduced here), switching is disallowed between auxiliaries
and main verbs, although the grammars of English and Spanish share the same phrase structure rules in the switched components. Still, such counterexamples notwithstanding, we
recognize the importance of Woolfords contribution in introducing codeswitching data into
linguistic theorizing (cf., also Woolford, 1984).
Also working within the generative model, Di Sciullo et al. (1986) posit that codeswitching is restricted by the Government Constraint, drawing on this X-bar-theoretical
hierarchical relation in disallowing codeswitching between particular elements in bilingual
speech (cf., also DIntrono, 1996).26 On this structural account, elements that stand in a
government relation (the governor and governee) must share the same language index, that
is, the government relation entails language coindexation. For example, the constraint predicts
that verbs and prepositions and their complements will be in the same language. But, this
prediction is contrary to what is in evidence in Spanish-English codeswitching: our participants accepted such switches in the reading task (21, drawn from 5), and they produced such
switches in the oral storytelling (22, drawn from 12) and written storytelling tasks (23, drawn
from 18).27

26

The Government Constraint, as stated in DiSciullo et al. (1986):


a.

If Lq carrier has index q, then Ymaxq

In a maximal projection Ymax, the Lq carrier is the lexical element which asymmetrically c-commands the other
lexical elements or terminal phrase nodes dominated by Ymaxq.
27
It merits pointing out that the systematic production of switching between prepositions and their complements in
participants writing samples confronts Joshis (1985) Constraint on Closed Items with a formidable empirical challenge; this, in addition to the theoretical argument leveled by Belazi et al. (1994), who suggest that the open/closed-class
distinction, an extragrammatical notion, should not impinge on the process of codeswitching, which is governed by
properly grammatical principles.
b.

The International Journal of Bilingualism

430

A. J. Toribio

(21) a. Al cumplir ella los veinte aos, el rey invit many neighboring princes to a party.
On her 20th birthday, the king invited/many neighboring princes to a party.
b. Since she was unmarried, he wanted her to choose un buen esposo.
Since she was unmarried, he wanted her to choose/a good husband.
(22) a. Blancanieves despert y los enanitos estaban very happy.
Snow White awoke and the darfs were / very happy.
b. She wanted to see la belleza que tena
She wanted to see/ the beauty that it had
(23) a. La madre de Caperucita le da a jar of honey.
Little Red Riding Hoods mother gave her/a jar of honey.
b. siempre traa a red cloth over her head.
always wore/a red cloth over her head.
c. para que le llevara food to her grandmother.
to take/food to her grandmother.
d. que su abuelita ocupaba some medicine for some sickness.
that her grandmother needed/some medicine for some sickness.
e. El lobo platica con Little Red Riding Hood for a while.
The wolf chatted with/Little Red RidingHood for a while.
f. goes to la casa de la abuelita
goes to/ the grandmothers house
g. and sent her to take honey to su abuela.
and sent her to make, to take honey to her/grandmother.
Therefore, while Di Sciullo et al. may be correct in proposing that codeswitching is constrained
by general principles that hold true of all natural languages, the formulation of this configurational constraint in terms of government is incorrect, as it proves overly restrictive, ruling
out permissible switches.
Continuing in this generativist vein in their investigation of codeswitching, Belazi et
al. (1994) argue that the coherence and co-occurrence restrictions attested in Spanish-English
codeswitching may be captured by reference to the Functional Head Constraint.28 In brief,
the proposal holds that a functional element and its complement will be drawn from the same
subclass of lexical items, precluding switching between functional elements such as modals,
auxiliaries, negation, determiners, and subordinating and coordinating conjunctions and
their complements.29 In the words of one informant, Algunas palabras dependen una de la

28

This constraint, grounded in the system of categories of Chomsky (1986) and the relations proposed in Abney (1987),
dictates that the semantic and syntactic features of a functional element must match the corresponding features of its
complement; the Functional Head Constraint merely extends the scope of f-selection to include language indexing.
Thus, like all other relevant features (e.g., finiteness and mood in subordinate clauses, and number and gender in
Noun Phrases), the [language] feature of a complement f-selected by a functional element, must match the corresponding [language] feature of that functional head.

29

The validity of the constraint has been supported with elicited imitation (Toribio et al. , 1993) and grammaticality judgment tasks (Toribio, 2000b).

The International Journal of Bilingualism

Accessing bilingual codeswitching competence

431

otra, y si estn divididas suena mal Some words depend on each other, and if they are
divided it sounds bad (Toribio, 2000b). Indeed, the predictions of the Functional Head
Constraint are strongly supported by the findings here. All of the bilingual participants
demonstrated a distinct pattern of responses to the text in which the same language was
maintained across a functional element and its complement, versus the text in which this relationship was compromised by a language switch. Recall that The Beggar Prince incorporated
switches at major category boundaries and between lexical elements and their complements
(e.g., between subject and predicate, between verb and object, between noun and subordinate clause, etc.), whereas Snow White and the Seven Dwarfs included switches that are
in violation of the constraint (e.g., between auxiliary and main verb, between object pronoun
and main or auxiliary verb, etc.). The findings are strengthened by participantsintrospection
on their judgments, which reveal a sensitivity to the dependency relation of f-selection.
Furthermore, the distinction between functional and lexical categories, which is implicated
in our bilingual participants reading performance, is also centrally implicated in their
codeswitched narrative productions.30 As amply illustrated in the storytelling task, speakers
failed to incorporate into their oral narratives codeswitching that would result in a violation
of the relation of f-selection.31 Thus, for example, there were attested no productions of
switching between auxiliary and main verb; and in those few instances in which their productions led to utterances that would have violated the Functional Head Constraint, various types
of disfluency, for example, lengthy pauses and self-repairs, served to circumvent the potentially offending switch. In contrast, switching in those contexts predicted to be
permissible between lexical elements (e.g., nouns, verbs, prepositions) and their complements, between subject and predicate, between adjunct modifiers and phrases modified was
smooth in oral performance and commonplace in the written samples (consider the numerous
and diverse examples of intrasentential codeswitching presented in (12) and (18) above).32
Therefore, though not its principle aim, these findings can be interpreted as presenting
forceful evidence of the extent to which the grammatical co-occurrence restrictions attested
in Spanish-English codeswitching adhere to the Functional Head Constraint.33

30

As elaborated by Belazi et al. 1994, the Functional Head Constraint is held to underlie linguistic competence; it is a
universal principle that is operative in monolingual and bilingual modes alike, though it finds additional, more visible
evidence in codeswitching.

31

One reviewer points out that determiner-noun pairings could be analyzed as counterexamples to the Functional Head
Constraint. As nouns are the most frequently borrowed category of words, it proves difficult to determine whether such
pairings are representative of insertional or alternational codeswitching, but consult Wentz and McClure (1976) and
Zentella (1981, 1997) who suggest that such alternations are ungrammatical as codeswitches.

32

There were, however, two ill-formed switches produced: switching between an indirect object clitic pronoun and a complex
verb (i) and switching after a coordinating conjunction (ii). This number of counterexamples is remarkable, given the
abundance of code-alternations attested in the study.
(i) Una ardilla que lo escucho le fue a tell a hunter and he went to look for the wolf. [sic]
A squirrel that heard him IND.OBJ.CL. went to/ tell
(ii) El hombre lobo beat Caperucita to abuelitas house and la asusto. [sic]
The wolf man/beat Little Red Riding Hood to/grandmothers/ house and/scared her.

33

For corroborating findings form second language bilingual codeswitching, consult Toribio (2000b).
The International Journal of Bilingualism

432

A. J. Toribio

5.2
Directions for further research

The findings of the present study also have implications for research in the cognate discipline
of psycholinguistics. For instance, the findings could serve to elucidate issues surrounding
bilingual speech processing (cf., De Bot, 1992; De Bot & Schreuder, 1993; Green, 1986;
Levelt, 1989; Poulisse, 1999; Poulisse & Bongaerts, 1994). In particular, the attested cooccurrence requirements support recent models which take account of the facts of bilingual
codeswitching by reference to language indexation, for example, the models advanced by
researchers such as Poulisse and Bongaerts (1994), who argue that words must contain information that specifies the language to which they belong (cf., Green, 1986), and even models
such as that proposed by De Bot (1992), in which the language-independent lexicon allows
for words of one language to form a subset (Paradis, 1987) that can be separately activated
(cf., De Bot & Schreuder, 1993). (Note that such models are consistent with the syntactictheoretical constructs previously discussed: the Government Constraint and the Functional
Head Constraint make reference to the labeling of lexical subclasses.34 )
One specific line of research that follows more directly from our findings relates to slips
of the tongue, as analyzed in Poulisse (1999). Basing her work on the well-founded assumption that such errors in native and second language speech are motivated by performance
variables (i.e., rather than problems with competence), Poulisse reasons that speakers should
be able to repair them. This is observed in our elicitation tasks: the ungrammatical codeswitching
sequences that gave rise to self-corrections demonstrated in the controlled reading task were
not evidenced in the oral narrative productions (and the ill-formed combinations elicited in
the reading task were not reproduced in the oral narrative or written narrative conditions).
These data may additionally be understood as corroborating Poulisses assertion that some
elements (e.g., function words and their complements) are commonly used and then stored
in combination and retrieved as units. However, it merits pointing out that while such units
may be less prone to error, their inviolability cannot be accredited solely to practice and
frequency, as Poulisse suggests. Recall that even those bilinguals who do not normally engage
in codeswitching as a social practice are nevertheless able to offer judgments on the wellformedness of specific combinations (cf., Toribio, 2000b).35 To be sure, much insight into
language processing is to be gained from further experimental investigation of codeswitched
speech production.
34

As argued in Toribio (2000b), these considerations allow for a reformulation of the Functional Head Constraint as making
reference to the labeling of lexical subclasses: a functional element and its complement will be activated from the same
subclass of lexical items. This informal restatement may be further articulated in terms of abstract feature matching
between functional structure and the lexical items that raise into them (cf., MacSwan, 1997; Rubin & Toribio, 1995;
Toribio & Rubin, 1996), and may alternatively be understood in view of Grimshaws (1991) notion of extended projections IP is an extended projection of VERB, DP is an extended projection of NOUN, etc as consistent with
Minimalist assumptions (Chomsky, 1993, 1995). Taking the bottom-up Minimalist perspective, a lexical category
creates the language domain for the projection. Still, irrespective of the specific analysis advocated for regulating
intrasentential codeswitching, it should be clear that differences between languages reduce to lexical properties, and
thus the patterns attested in bilingual productions must derive from the interaction between the two component lexicons (cf., MacSwan, 1997; Muysken, 1995).

35

This disparity could well be due to differences in the populations under study: Poulisses interest is in the language
productions of foreign language learners (as compared with native speakers), while the present study concerns the linguistic
behavior of more fluent bilinguals.

The International Journal of Bilingualism

Accessing bilingual codeswitching competence

433

5.3
Conclusion

To conclude, the present work has established several methodologies that are valid and informative in the study of the linguistic competence that underlies language alternations in
Spanish-English bilingual speech. The reliability of the reading, recounting, and writing
tasks is affirmed, as revealed by the uniform behaviors of the bilingual participants tested;
in addition, the reliability of the data is confirmed by the recurrent patterns observed across
tasks. Finally, the data were subjected to preliminary analysis, confronting and advancing
syntactic theoretical proposals, and pointing to directions for further research into bilingual
language processing.
Received: February, 2000; revised: October, 2000; accepted: January, 2001

References
ABNEY, S. (1987). The English noun phrase in its sentential aspect. Doctoral dissertation, Massachusetts
Institute of Technology, Cambridge, MA.
AGUIRRE, A., Jr., (1977). Acceptability judgment of grammatical and ungrammatical forms of intrasentential code alternation. Doctoral dissertation, Stanford University, Stanford, CA.
AUER, P. (1998). Codeswitching in conversation. New York: Routledge.
BELAZI, H., RUBIN, E., & TORIBIO, A. J. (1994). Codeswitching and X-Bar Theory: The Functional
Head Constraint. Linguistic Inquiry, 25, 221 237.
BHATIA, T., & RITCHIE, W. (1996). Bilingual language mixing, universal grammar, and second
language acquisition. In W. Ritchie & T. Bhatia (Eds.), The handbook of second language acquisition (pp.627 688). New York: Academic Press.
BIRDSONG, D. (1989). Metalinguistic performance and interlinguistic competence. New York: Springer
Verlag.
CHOMSKY, N. (1981). Lectures on government and binding. Dordrecht: Foris.
CHOMSKY, N. (1986). Barriers. Cambridge, MA: MIT Press.
CHOMSKY, N. (1991). Some notes on economy of derivation and representation. In R. Friedin (Ed.),
Principles and parameters in comparative grammar (pp.417 454). Cambridge, MA: MIT Press.
CHOMSKY, N. (1993). A Minimalist Program for linguistic theory. In K. Hale & S. J. Keyser (Eds.),
The view from building 20 (pp.1 52). Cambridge, MA: MIT Press.
CHOMSKY, N. (1995). The Minimalist Program. Cambridge, MA: MIT Press.
De BOT, K. (1992). A bilingual production model: Levelts speaking model adapted. Applied Linguistics,
12, 1 24.
De BOT, K., & SCHREUDER, R. (1993). Word production and the bilingual lexicon. In R. Schreuder
& B. Weltens (Eds.), The bilingual lexicon (pp. 191 214). Amsterdam: John Benjamins.
DINTRONO, F. (1996). Spanish-English codeswitching conditions on movement. In J. Jensen & A.
Roca (Eds.), Spanish in contact (pp.187 201). Somerville, MA: Cascadilla Press.
Di SCIULLO A-M., MUYSKEN, P., & SINGH, R. (1986). Government and codemixing. Journal of
Linguistics, 22, 1 24.
The International Journal of Bilingualism

434

A. J. Toribio

GINGRS, R. (1974). Problems in the description of Spanish/English intrasentential codeswitching.


In G. Bills (Ed.), Southwest areal linguistics(pp.167 174). San Diego, CA: University of California
Institute for Cultural Pluralism.
GREEN, D. (1986). Control, activation and resource: A framework and a model for the control of speech
in bilinguals. Brain and Language, 27, 210 233.
GRIMSHAW, J. (1991). Extended projections. Unpublished manuscript, Brandeis University.
GROSJEAN, F. (1998). Studying bilinguals: Methodological and conceptual issues. Bilingualism:
Language and Cognition, 11, 131 149.
GUMPERZ, J. (1971). Language in social groups. Stanford, CA: Stanford University Press.
GUMPERZ, J. (1976). The sociolinguistic significance of conversational codeswitching. Working papers
of the language behavior research laboratory, #46. Berkeley, CA: University of California.
GUMPERZ, J. (1982). Discourse strategies. Cambridge, U.K.: Cambridge University Press.
GUMPERZ, J., & TORIBIO, A. J. (1999). Codeswitching. In F. Keil & R. Wilson (Eds.), The MIT encyclopedia of the cognitive sciences (pp.118 119). Cambridge, MA: MIT Press.
JACOBSON, R. (1977). The social implications of intrasentential codeswitching. The New Scholar,
6, 227 256.
JACOBSON, R. (1990). Codeswitching as a worldwide phenomenon. New York: Peter Lang.
JAKOBSON, R., FANT, G. M., & HALL, M. (1952). Preliminaries to speech analysis: The distinctive
features and their correlates. Cambridge, MA: The MIT Press.
JOSHI, A. (1985). Processing of sentences with intrasentential codeswitching. In D. Dowty, L. Karttunen
& A. Zwicky (Eds.), Natural language parsing: Psychological, computational and theoretical
perspectives(pp.290 205). Cambridge, U.K.: Cambridge University Press.
KELLER, G. (1979). The literary stratagems available to the bilingual Chicano writer. In F. Jimnez (Ed.),
The identification and analysis of Chicano literature(pp.262 316). Ypsilanti, MI: Bilingual Press.
LANCE, D. (1975). Spanish/English codeswitching. In E. Hernndez-Chvez, A. Cohen & A. Beltramo
(Eds.), El lenguaje de los Chicanos (pp.138 153). Arlington, VA: Center for Applied Linguistics.
LEVELT, W. (1989). Speaking: From intention to articulation. Cambridge, MA: MIT Press.
LIPSKI, J. (1978). Codeswitching and the problem of bilingual competence. In M. Paradis (Ed.), Aspects
of bilingualism (pp.250 264). Columbia, SC: Hornbeam Press.
LIPSKI, J. (1985). Linguistic aspects of Spanish-English language switching. Tempe, AZ: Arizona State
University, Center for Latin American Studies.
MacSWAN, J. (1997). A Minimalist approach to intrasentential codeswitching:Spanish-Nahuatlbilingualism
in Central Mexico. Doctoral dissertation, University of California, Los Angeles.
MAHOOTIAN, S., & SANTORINI, B. (1996). Codeswitching and the complement/adjunct distinction. Linguistic Inquiry, 27, 464 479.
McCLURE, E. (1977). Aspects of codeswitching in discourse of bilingual Mexican-American children.
In M. Saville-Troike (Ed.), Linguistics and anthropology(pp.93115). Washington, DC: Georgetown
University Press.
McCLURE, E. (1981). Formal and functional aspects of the codeswitched discourse of bilingual children. In R. Duran (Ed.), Latino language and communicative behavior(pp.69 94). Norwood, NJ:
Ablex Publishing.
MEISEL, J. (1989). Early differentiation of languages in bilingual children. In K. Hyltenstam & L.
Obler (Eds.), Bilingualism across the lifespan: Aspects of acquisition, maturity, and loss (pp.13 40).
Cambridge, U.K.: Cambridge University Press.
MEISEL, J. (1994). Codeswitching in young bilingual children: The acquisition of grammatical constraints.
Studies in Second Language Acquisition, 16, 413 439.
The International Journal of Bilingualism

Accessing bilingual codeswitching competence

435

MILROY, L., & MUYSKEN, P. (1995). One speaker, two languages. Cambridge, U.K.: Cambridge
University Press.
MONTES-ALCAL, C. (2000). Two languages, one pen: Sociopragmatic functions in written SpanishEnglish codeswitching. Doctoral dissertation, University of California, Santa Barbara.
MUYSKEN, P. (1995). Codeswitching and grammatical theory. In L. Milroy & P. Muysken (Eds.), One
speaker, two languages (pp.177 198). Cambridge, U.K.: Cambridge University Press.
MYERS-SCOTTON, C. (1993). Social motivations of codeswitching. Oxford: Clarendon Press.
NISHIMURA, M. (1997). Japanese/English code-switching:Syntax and pragmatics. New York: Peter Lang.
PEALOSA, F. (1980). Chicano sociolinguistics.Rowley, MA: Newbury House.
PARADIS, M. (1987). The assessment of bilingual aphasia. Hillsdale, NJ: Lawrence Erlbaum Associates.
PFAFF, C. (1976). Functional and syntactic constraints on syntactic variation in code-mixing. In S.
Steever, C. Walker, & S Mufwene (Eds.), Papers from the parasession on diachronic syntax (pp.
248 259). Chicago Linguistics Society.
PFAFF, C. (1979). Constraints on language mixing: Intrasentential codeswitching and borrowing in
Spanish/English. Language, 55, 291 318.
POPLACK, S. (1980). Sometimes Ill start a sentence in English Y TERMINO EN ESPAOL: Toward a
typology of codeswitching. In J. Amastae & L. Elas-Olivares (Eds.), Spanish in the United States:
Sociolinguistic aspects (pp.230 263). Cambridge, U.K.: Cambridge University Press.
POPLACK, S. (1981). Syntactic structure and social function of codeswitching. In R. Duran (Ed.),
Latino language and communicative behavior (pp.169 184). Norwood, NJ: Ablex Publishing.
POPLACK, S. (1983). Bilingual competence: Linguistic interference or grammatical integrity? In L. ElasOlivares (Ed.), Spanish in the U.S. setting: Beyond the Southwest (pp. 107 129). Rosslyn, VA:
National Clearinghouse for Bilingual Education.
POULISSE, N. (1999). Slips of the tongue: Speech errors in first and second language production.
Amsterdam: Benjamins.
POULISSE, N., & BONGAERTS, T. (1994). First language use in second language production. Applied
Linguistics, 15, 36 57.
RAKOWSKY, A. (1989). A study of intrasentential codeswitching in Spanish-English bilinguals and
second language learners. Doctoral dissertation, Brown University, Providence, Rhode Island.
ROMAINE, S. (1995). Bilingualism. Cambridge, MA: Blackwell.
RUBIN, E., & TORIBIO, A. J. (1995). Feature checking and the syntax of language contact. In J. Amastae,
G. Goodall, M. Montalbetti & M. Phinney (Eds.), Contemporary research in Romance linguistics
(pp.177 185). Amsterdam: Benjamins.
SCHTZE, C. (1996). The empirical base of linguistics: Grammaticality judgments and linguistic methodology. Chicago: Chicago University Press.
TIMM, L. A. (1975). Spanish-English codeswitching. El porque and how-not-to. Romance Philology,
28, 473 482.
TORIBIO, A. J. (2000a). Spanglish? ! Bite your tongue! Spanish-English codeswitching among U.S. Latinos.
In R. Flores (Ed.), Reflexiones, 1999 (pp.115 147). Austin, TX: Center for Mexican American
Studies.
TORIBIO, A. J. (2000b). On the emergence of codeswitching competence. Forthcoming in Bilingualism:
Language and Cognition.
TORIBIO, A. J. (2000c). Constraints on operator movement in English-Spanish codeswitching: Generative
perspectives, from Government-Binding to Minimalism. Under review, for J. MacSwan (Ed.),
Grammatical theory and bilingual codeswitching,. Cambridge, MA: MIT Press.
The International Journal of Bilingualism

436

A. J. Toribio

TORIBIO, A. J. (2000d). Codeswitching and minority language attrition. Forthcoming in R. Leow and
C. Sanz (Eds.) L1 and L2 acquisition of Spanish. Washington, DC: Georgetown University Press.
TORIBIO, A. J. (2000e). Once upon a time en un lugar muy lejano Spanish-English codeswitching
patterns across narratives. Forthcoming in A. Roca (Ed.), Spanish in the United States: Linguistic
issues and challenges. Somerville, MA: Cascadilla Press.
TORIBIO, A. J. (2000f). Verb raising and verbal morphology in Spanish language attrition. Paper presented
the 25th Annual Boston University Conference on Language Development, Boston, MA.
TORIBIO, A. J., & RUBIN, E. (1996). Codeswitching in generative grammar. In J. Jensen & A. Roca
(Eds.), Spanish in contact (pp.203 226). Somerville, MA: Cascadilla Press.
TORIBIO, A. J., ROEBUCK, R., LANTOLF, J., & PERRONE, A. (1993). Syntactic constraints on codeswitching: Evidence of abstract knowledge in second language acquisition. Unpublished manuscript,
Cornell University.
TORRES, L. (1997). Puerto Rican discourse: A sociolinguistic study of a New York suburb. Mahwah, NJ:
Lawrence Erlbaum Associates.
VALDS, G. (1976). Social interaction and codeswitching patterns. In G. D. Keller, R. V. Teschner &
S. Viera (Eds.), Bilingualism in the bicentennial and beyond (pp.52 85). Jamaica, NY: Bilingual
Press.
WENTZ, J., & McCLURE, E. (1976). Aspects of the syntax of the codeswitched discourse of bilingual
children. In F. Ingemann (Ed.), 1975 Mid-America Linguistics Conference papers (pp. 351 370).
Lawrence, KS: University of Kansas. Washington, DC: Educational Resources Information Center.
WHITE, L. (1996). Universal Grammar and second language acquisition: Current trends and new directions. In W. Ritchie & T. Bhatia (Eds.), The handbook of second language acquisition(pp.85 120).
New York: Academic Press.
WOOLFORD, E. (1983). Bilingual codeswitching and syntactic theory. Linguistic Inquiry, 14, 520 536.
WOOLFORD, E. (1984). On the application of wh-movement and inversion in codeswitching sentences.
Revue Qubcoise de Linguistique, 14, 77 86.
ZABALETA, F. (2000). Aspectos morfosintcticos del espaol como lengua materna de universitarios
californianos. Spanish-English codeswitching patterns across narratives. Forthcoming in A. Roca
(Ed.), Spanish in the United States: Linguistic issues and challenge. Somerville, MA: Cascadilla
Press.
ZENTELLA, A. C. (1981). Hablamos los dos. We speak both: Growing up bilingual in el barrio. Doctoral
dissertation, University of Pennsylvania, Philadelphia.
ZENTELLA, A. C. (1997). Growing up bilingual. Malden, MA: Blackwell Publishers.

The International Journal of Bilingualism